Statistically distinguishable rating scale

The article proposes a method of designing a statistically distinguishable rating scale that is not excessive in relation to the existing observation statistics. This allows for more stable validation with a fixed maximum number of violations of the Wald criterion compared to an excess scale, which is usually used by banks. The increased robustness of validation will reduce the calibration probability of default, which provides savings in capital requirements under the advanced IRB approach. Theoretical justifications of the effect are presented, numerical calculations for three rating scales, two of which are open statistics of rating agencies, and the third is the rating scale of one bank from closed data are performed. The proposed method is most relevant for the corporate segment of the loan portfolio.